To understand how we perceive color, it is necessary to learn how our retina creates visual signals from the incoming flux of light. It is known that 3 types of cone photoreceptors are the starting point for any color signals. The cone responses to light are passed to retinal ganglion cells, and their axons leave the eye to provide input to the thalamus, which in turn sends signals to primary visual cortex (V1). Currently, there are two persistent controversies over the chromatic structure of the receptive fields of neurons along this pathway. One controversy concerns the midget class of retinal ganglion cells near the fovea. It is uncertain if these ganglion cells receive input from only single cone types, or from mixed cone types, in their receptive field centers. The two options lead to different color coding schemes at this stage of the visual system, and thus constrain how color is processed at later stages. Because the midget ganglion cells comprise about 80% of the output from the retina, it is crucial to work out their true signaling properties. The second controversy centers on color signaling in V1. Here, it has also been unclear how sensitive are V1 neurons to cone-specific stimuli, and the difficulty has been, in part, inherited from problems associated with uncertainties in the retinal input. The main impediments to solving these problems have been the inability to identify and stimulate individual cones in the retina in vivo. The goal of this proposal is to employ a newly developed retinal microscope that overcomes these hurdles. The instrument can image the cones in a living eye, and can stimulate single cones selectively and repeatably with colored lights. We first propose to verify that cones can be mapped by their spectral sensitivity in humans, using psychophysical techniques. Using the stimulation parameters that allow cone identities to be obtained, we next propose to map physiologically the cone fields that provide input to single neurons in the visual thalamus of a trichromatic primate. With cone-sized stimulation, thalamic neurons receive input essentially from single retinal ganglion cells; thus we will learn whether ganglion cell receptive field centers are composed of pure or mixed cone types. We will focus on foveal thalamic neurons that receive input from the midget retinal ganglion cell class. Finally, we propose to map the cone fields of V1 neurons, to determine the strength of their cone-specific input. We will confirm histologically the location of these neurons in visual cortex, to learn where they are situated within the known anatomical circuits of V1. Our results will afford the first direct mapping of cone fields in vivo, and improve our understanding of how photoreceptor signals are processed by neurons subserving foveal vision.